Final version of 07/12/2011 EUROPEAN COMMISSION DG Regional Policy DG Employment, Social Affairs and Equal Opportunities COCOF_11-0041-01-EN GUIDANCE ON TREATMENT OF ERRORS DISCLOSED IN THE ANNUAL CONTROL REPORTS (Annex to the Guidance on ACRs and Opinions of 18/02/2009, ref. COCOF 09/0004/01- EN) This is a Working Document prepared by the Commission services. On the basis of the applicable Community law, it provides technical guidance to the attention of public authorities, practitioners, beneficiaries or potential beneficiaries, and other bodies involved in the monitoring, control or implementation of Cohesion Policy on how to interpret and apply the Community rules in this area. The aim of the working document is to provide Commission services' explanations and interpretations of the said rules in order to facilitate the implementation of operational programmes and to encourage good practices. However, this guidance is without prejudice to the interpretation of the Court of Justice and the Court of First Instance or evolving Commission decision making practice.
Final version of 07/12/2011 COCOF_11-0041-01-EN List of contents LIST OF ACRONYMS...4 GLOSSARY...5 1. INTRODUCTION...7 2. EVALUATION OF ERRORS...8 2.1. Approach set out in previous guidance from the Commission...8 2.2. Systemic errors...8 2.3. Random errors...9 2.4. Anomalous errors...9 2.5. Errors relating to unfinished contradictory procedure...10 2.6. Total projected error rate...11 2.7. Total projected error rate and assessment of management and control systems 12 3. DISCLOSURE OF ERROR RATES IN ACR VIA SFC 2007...13 4. INADEQUATE AUDIT OPINIONS...13 5. CORRECTIVE MEASURES...14 5.1. Concept of corrected error for determining the audit opinion...14 5.2. Corrective measures as subsequent events...14 5.3. Option 1: extrapolated financial correction...15 5.4. Option 2: Correction of each type of error...16 5.5. Net off with an expenditure "buffer"...16 6. IMPLICATIONS FOR SAMPLING RESULTING FROM HIGH ERROR RATE...17 7. MOST LIKELY ERROR AND UPPER ERROR LIMIT...18 8. PARTICULAR CASES...19 8.1. Errors detected by AA in expenditure that was also considered irregular by the Managing Authority, Intermediate Body or Certifying Authority...19 8.1.1. Irregularities already detected and acted upon by the IB/MA/CA, but not yet corrected before the sample was drawn by AA...19 8.1.2. Irregularities detected during controls by IB/MA and corrected insufficiently before the sample was drawn by the AA...20
8.1.3. Irregularities relating to expenditure "de-certified" after the sample was drawn by the AA...20 8.2. Net off overstatement errors against understatement errors to arrive at an overall most likely error...21 ANNEX 1 TOTAL PROJECTED ERROR RATE...22 ANNEX 2 - TABLE FOR DECLARED EXPENDITURE AND SAMPLE AUDITS...23 3/23
Final version of 07/12/2011 COCOF_11-0041-01-EN LIST OF ACRONYMS AA Audit Authority AAR Annual Activity Report (of each Directorate-General of the Commission) ACR Annual Control Report CA Certifying Authority CF Cohesion Fund COCOF Committee of the Coordination of the Funds DAS Declaration of Assurance from the European Court of Auditors on the implementation of the EU budget ERDF European Regional Development Fund ESF European Social Fund IB Intermediate Body ISA - International Standard of Auditing MA Managing Authority MCS - Management and Control System MS Member State MUS - Monetary Unit Sampling
GLOSSARY Term Anomalous error Contradictory procedure Error Expenditure of year N Irregularity Known error Definition A misstatement that is demonstrably not representative of the population. Procedure whereby (draft) audit reports are sent to the auditee with a request for a written reply within a given time-limit. For the purposes of this guidance, an error is a quantifiable overstatement of the expenditure certified declared to the Commission. Expenditure declared to the Commission, on the basis of which the sample of operations is selected. Same meaning as error. A known error is an error found outside the sample audited. In case of MUS sampling, a known error is also the amount of error detected in a sample item with a value equal or exceeding the value of the selection interval Misstatements Population Random error Sample error rate Same meaning as error. The entire set of data from which the sample is selected (for the purposes of Article 62.1(b) of Regulation (EC) N 1830/2006) and about which the auditor wishes to draw conclusions. The errors which are not considered systemic are classified as random errors. This concept presumes the probability that random errors found in the audited sample are also present in the non-audited population. The sample error corresponds to the amount of irregularities detected (by the audits of operations (carried out under Article 62.1(b) of Regulation (EC) N 1830/2006) divided by the expenditure 5/23
Term Definition audited. Systemic error Total projected error rate The systemic errors are errors found in the sample audited that have an impact in the non-audited population and occur in well defined and similar circumstances. These errors generally have a common feature, e.g. type of operation, location or period of time. They are in general associated with ineffective control procedures within (part of) the management and control systems. The total projected error corresponds to the sum of the following errors: projected random errors, systemic errors, known errors and anomalous errors not corrected. The AA should compare the total projected error rate with the materiality threshold in order to reach conclusions for the total population covered by the sample. 6/23
1. INTRODUCTION The present document aims to provide additional guidance by clarifying the main questions raised by Member States (MS) in relation to the error rates disclosed in the Annual Control Report (ACR) and the impact of those error rates in the audit opinion 1 of the Audit Authority (AA). The guidance is a joint document prepared by the Directorate-General of Regional Policy in cooperation with the Directorate-General for Employment, Social Affairs and Equal Opportunities. As such, the guidance is to be applied by AA responsible for providing an ACR by end 2011 onwards for programmes co-financed by ERDF, CF or ESF. The draft guidance does not bring any new or additional tasks to the national authorities. The concepts explained below intend to clarify the questions raised by the AAs in the context of the assessment of the ACRs submitted by end 2010 and questions raised by the AA during the draft versions of the present guidance discussed in the Homologues Group meeting of 17-18 October 2011 and in the COCOF meeting of 23 November 2011. The guidance's sole intention is to clarify the approach that the AAs should have when analysing the results of the audits of operations and systems audits. This guidance does not replace existing Commission guidelines, namely the following: Guidance on ACRs and Opinions of 18/02/2009, ref. COCOF 09/0004/01-EN, herafter "guidance on ACRs and Opinions"; Guidance note on sampling methods for Audit Authorities of 15/09/2008, ref. COCOF 08/0021/01-EN, hereafter "guidance on sampling"; Guidance on a common methodology for the assessment of management and control systems [MCS] in the Member States" 2, ref. COCOF 08/0019/00-EN, hereafter "guidance for the assessment of MCS"; Guidance note to Certifying Authorities on reporting on withdrawn amounts, recovered amounts, amounts to be recovered and amounts considered irrecoverable, applicable to programming period 2007-2013 and the remainder of 1 2 Article 62(1)(d)(ii) of Regulation (EC) No 1083/2006 requires the audit authority to provide an opinion as to whether the management and control system functions effectively, so as to provide reasonable assurance that the statements of expenditure presented to the Commission are correct and, as a consequence, reasonable assurance that the underlying transactions are legal and regular. See COCOF note 08/0019/00-EN, in which four categories for the assessment of the systems are foreseen: Category 1: Works well; only minor improvements needed (high reliability), category 2: works but some improvements are needed (average reliability), category 3: works partially; substantially improvements are needed (average reliability), category 4: essentially does not work (low reliability). 7/23
the 2000-2006 programming period, of 27/03/2010, ref. COCOF N 10/0002/00/EN, hereafter "guidance to CA". 2. EVALUATION OF ERRORS 2.1. Approach set out in previous guidance from the Commission As set out in section 5 (last bullet point) of the guidance on ACRs and Opinions and on section 6.8 of the guidance note on sampling, the ACR should present the evaluation of errors detected, in addition to the correction of these errors. In the mentioned section 6.8 the Commission also refers to International Standard of Auditing (ISA) n 530 3, according to which the auditor should consider the sample results, the nature and cause of any errors identified, and their possible effect on the particular audit objective and on other areas of the audit. Hence, since at least 2008 the guidance provided by the Commission made clear that the AA is expected that they perform a qualitative in-depth analysis of the errors detected in the audits of operations and that such analysis is presented in the ACR. As such, in the section of the ACR dealing with the audits on operations, the AA should explain the nature of the errors which have impact upon the total projected error rate, as errors may arise from, inter alia, public procurement, financial engineering instruments and aid schemes. As set out in the guidance on ACRs and Opinions of 2009 4, the ACR should indicate whether any problems (irregularities) identified were considered to be of a systemic character, and of the measures taken, including a quantification of the irregular expenditure and any related financial corrections. In any case, this guidance cannot replace the professional judgment of the AA when assessing whether an error is systemic, random or anomalous. This assessment is necessarily a case-by-case analysis, to be set out in the ACR. 2.2. Systemic errors The systemic errors are errors found in the sample audited that have an impact in the non-audited population and occur in well defined and similar circumstances. They are in general associated with ineffective control procedures within (part of) the management and control systems. Indeed, the identification of a potential systemic error implies carrying out the complementary work necessary for the identification of its total extent and subsequent quantification. This means that all the situations susceptible of containing an error of the same type as the one detected in the sample 3 4 http://web.ifac.org/download/a027-2010-iaasb-handbook-isa-530.pdf See section 4, last bullet point of that guidance document. 8/23
should be identified, thus allowing the delimitation of its total effect in the population 5. According to Article 98(4) of regulation (EC) N 1083/2006, "in the case of a systemic irregularity, the Member State shall extend its enquiries to cover all operations liable to be affected". Hence, the concept of systemic irregularity and the measures the Member State needs to take are well known to all actors of the MCS. Similarly to systemic errors, it can be that an error found in the sample leads the auditor to detect one or more errors outside that sample this can be classified as "known errors". For example, if a contract is found to be illegal under the public procurement rules it is likely that part of the related irregular expenditure has been declared in a payment claim or invoice included in the sample audited and the remaining expenditure has been declared in payment claims or invoices not included in that sample. The treatment of known errors is the same as for systemic errors. 2.3. Random errors The errors which are not considered systemic are classified as random errors. This concept presumes the probability that random errors found in the audited sample are also present in the non-audited population, since the sample is representative. Hence, these errors are to be included in the calculation of the projection of errors see section 2.6 of this guidance. 2.4. Anomalous errors In the ACRs submitted by the end of 2010, there were many cases reported as anomalous errors by the AA, resulting in projected error rates lower than should have been reported. A statistical sample is representative for the population and therefore anomalous errors should only be accepted in very exceptional, well motivated circumstances. The frequent recourse to this concept without a due justification may undermine the reliability of the audit opinion. The AA is required to provide in the ACR a high degree of certainty that such an anomalous error is not representative of the population and to explain the additional audit procedures it carried out to conclude on the existence of an anomalous error, as required by the ISA n 530. The ISA n 530 further specifies: "A.19. When a misstatement has been established as an anomaly, it may be excluded when projecting misstatements to the population. However, the 5 For example, it can be that a certain error has been detected in an operation co-financed under a priority axis relating to financial engineering. It may be that this error occurs in other operations in the same priority axis. The AA needs to determine if this is the case, in cooperation with the MA/IB. But, as already stated, any error found in the sample is to be included in the calculation of the projected error rate (except in duly justified anomalous errors), for the simple reason that the sample is representative of the population. 9/23
effect of any such misstatement, if uncorrected, still needs to be considered in addition to the projection of the non-anomalous misstatements". A.22. In the case of tests of details, the projected misstatement plus anomalous misstatement, if any, is the auditor s best estimate of misstatement in the population. When the projected misstatement plus anomalous misstatement, if any, exceeds tolerable misstatement, the sample does not provide a reasonable basis for conclusions about the population that has been tested. ( )" This means that, when the AA decides to exclude an anomalous error from the calculation of the projected error, the amount of the anomalous error is to be added in the calculation of the total projected error rate if it has not been corrected, in accordance with section 5.1 of this guidance. If the anomalous error has been corrected then it does not count for the total projected error rate. This approach is only applicable to anomalous errors given their exceptional nature, already foreseen in the quoted audit standard. 2.5. Errors relating to unfinished contradictory procedure The errors considered in the total projected error rate should relate to findings disclosed in a final audit report, i.e. after the contradictory procedure with the auditee has been concluded. In duly justified cases where such contradictory procedure was not concluded before submission of the ACR, this could constitute a limitation in scope and a qualified opinion may be provided on the basis of the AA's professional judgement. In this case, the AA should state in the audit opinion whether such limitations have an impact on the declared expenditure, and if so providing quantification 6. Depending on the stage of the contradictory procedure of a given audit, it can happen that the AA can either: (i) consider the errors found in that audit in the calculation of the projection of errors (as defined in section 2.6 of this guidance) or (ii) quantify the irregularities in a scope limitation to the audit opinion, together with an indication of whether the total error rate would be materially affected by that potential irregularity (i.e. if the total projected error including this irregularity would result in a total error rate above 2%). In any case, the following is still valid: "The error rate to be provided in the annual control report will normally be based on the final audit results (after the contradictory procedure) related to the sample selected for the reference period. Nonetheless, it could happen that following further follow up in line with the administrative/audit procedures, it might be concluded that an error is finally not considered as being an error. As a consequence, the error rate indicated in the previous reports may need to be updated." 7 6 7 See section 7 of the guidance on ACRs and Opinions. See footnote 9 of the guidance on ACRs and Opinions. 10/23
2.6. Total projected error rate The AA should also disclose in the ACR the total projected error rate, which the AA should compare with the materiality threshold in order to reach conclusions for the population, as follows from the second subparagraph of Article 17(4) of Commission Regulation (EC) No 1828/2006. According to the 2 nd paragraph of this provision, "in operational programmes for which the projected error rate is above the materiality level, the audit authority shall analyse its significance and take the necessary actions, including making appropriate recommendations, which will be communicated in the annual control report". The total projected error rate represents the estimated effect of the errors in the management and control systems, in percentage of the population for the year N. The total projected error should reflect the analysis done by the AA in regard to the errors detected in the context of the audits of operations carried out under Article 62.1(b) of Regulation (EC) N 1830/2006. The total projected error corresponds to the sum of the following errors: projected random errors, systemic errors and uncorrected anomalous errors see flowchart in Annex 1 of this guidance. If systemic errors are identified in the audited sample and their extension in the population not audited is delimited with precision, then the systemic errors relating to the population are added to the total projected error. If such delimitation is not done before the ACR is submitted, the systemic errors are to be treated as random for the purposes of the calculation of the projected random error. Concerning random errors, the calculation of the projection of errors differs according to the sampling method selected and described in the audit strategy. For projection of errors, see sections 6.3 to 6.6 of the Commission's guidance on sampling methods examples of sampling methods and their application. In general terms: If the AA uses the Monetary Unit Sampling (MUS) as its sampling method, the projected random error corresponds to the most likely error (MLE). Similarly, for difference estimation, the AA should calculate the achieved precision and indicate the upper and lower limit levels thus obtained. For non-statistical sampling, the projection of the error rate is required under the formal approach explained in section 6.6 of the guidance on sampling methods, where such approach is applied. If for non-statistical sampling the method used is not the mentioned formal approach, there is no extrapolation of the error rate, i.e. the projected error rate is the error rate of the sample All errors should be quantified by the AA and included in the total projected error rate, with the exception foreseen in section 2.4 of this guidance. Without this 11/23
quantification the error rate cannot be considered reliable since it is probably understated. In this circumstance, the audit opinion should be qualified. In general, all errors found are to be taken into account for calculation of the total projected error rate. Section 8 of the present document refers to particular cases where this may not be the case. 2.7. Total projected error rate and assessment of management and control systems If the total projected error rate is above the materiality level of 2%, this is an indication that the declared expenditure is irregular to a material extent, i.e. the Management and Control System (MCS) has not functioned properly. Section 2 of Annex IV to Regulation (EC) No 1828/2006 requires the AA to combine the results of systems audits and audit of operations in order to obtain a high level of assurance on the effective functioning of the MCS. In line with the guidance for the assessment of MCS 8, the AA should, on the basis of the system audits carried out (and taking into consideration any compensating controls that may exist), formulate an overall conclusion by system, which will contribute to the AA's audit opinion. Section 6 of the guidance on ACRs and Opinions provides already some cases to be considered by the AA when assessing the overall assurance resulting from systems audits and audits of operations. If the AA considers that the MCS is in category 2 and the total projected error rate is below the materiality level of 2%, the audit opinion may be unqualified. However, if the MCS is classified in category 1 or 2 and the total projected error rate is above 2% this indicates that, despite the relatively positive assessment resulting from the systems audits carried out by the AA, the MCS is in practice not sufficiently effective in preventing, detecting and correcting irregularities and recovering amounts unduly paid. A qualified audit opinion is therefore deemed appropriate in this case. If the AA considers that the MCS is in category 3 or 4, a qualified audit opinion is deemed appropriate 9, even if the total projected error rate is below 2%. The AA should use its professional judgment to assess whether the systems audits that led to the classification of the MCS in category 3 are based on findings that could not have been detected by the audits of operations (e.g. systems audits covering MCS that were modified after year N, thus reducing the assurance granted by the audits of operations for the expenditure in that year). 8 9 This guidance refers to four categories for the assessment of the systems: Category 1: Works well; only minor improvements needed (high reliability), category 2: works but some improvements are needed (average reliability), category 3: works partially; substantially improvements are needed (average reliability), category 4: essentially does not work (low reliability). The expression "deemed appropriate" implies that the professional judgment of the AA is required in order to drawn appropriate conclusions on its work. 12/23
If the MCS is in category 3 or 4 and the total projected error rate is above 2% then the opinion is qualified or adverse. The differentiation between qualified and adverse opinion depends on the severity of the errors and of the magnitude of the error. 3. DISCLOSURE OF ERROR RATES IN ACR VIA SFC 2007 The aforementioned information should be presented in the ACR section concerning audits on sample of operations. In addition, the "table for declared expenditure and sample audits" to be attached to the ACR (as follows from table 9 of Annex VIII of regulation (EC) N 1828/2006) should disclose the error rate found in the sample and the total projected error rate (as defined under section 2.6) see Annex 2to the present document. The ACR is to be submitted to the Commission via SFC2007. The module in SFC2007 includes the above-mentioned table, which will have to be filled in by the AA. The information on the total projected error rate is to be inserted in a separate column, after the column entitled "amount and percentage (error rate) of irregular expenditure in random sample". The existing table in SFC2007 has been modified in view of the submission of the 2011 ACRs, in order to solve technical issues raised during the last submission of ACRs. While it is methodological correct to report one error rate covering the programmes included in a common MCS, it may not be always true that the opinion is the same for all the programmes of that system. Article 62(1)d(ii) of Council regulation (EC) No 1083/2006 refers that "When a common system applies to several operational programmes, the information referred to in point (i) may be grouped in a single report, and the opinion and declaration issued under points (ii) and (iii) may cover all the operational programmes concerned". However, if systems audits or the analysis of the errors detected in the common sample show particular deficiencies for one single programme under the common MCS, the AA may consider differentiating its audit opinion for this particular programme. As such the SFC2007 allows the AA to insert different opinions for each programme, even if they belong to the same MCS. 4. INADEQUATE AUDIT OPINIONS On the basis of the experience obtained from the submission of the 2010 ACRs, the Commission considers the following cases as inadequate audit opinions: unqualified opinions although no audits of operations on the year N expenditure were carried out; unqualified opinion even though the AA has not audited all the operations in the sample; 13/23
unqualified opinions although the total projected error rates were above the materiality level, and/or significant weaknesses had been detected in the system audits, without the appropriate corrective measures being taken by the national authorities in time before the disclosure of the audit opinion; disclaimer of opinion because the contradictory procedures for audits of operations were not finalised. 5. CORRECTIVE MEASURES 5.1. Concept of corrected error for determining the audit opinion All errors detected in the sample audited by the AA have to be notified in accordance to the provisions of article 28 of (EC) Regulation 1828/2006 as amended by (EC) Regulation 846/2009, corrected and reported according to the guidance on sampling. Pursuant to Article 70 of Regulation (EC) No 1083/2006 and Article 70 of Council Regulation (EC) No 1198/2006, Member States are required to correct and recover amounts unduly paid. Member States have two choices: 1) withdrawing the irregular expenditure from the programme immediately when they detect the irregularity, by deducting it from the next statement of expenditure, thereby releasing EU funding for commitment to other operations or 2) leaving the expenditure for the time being in the programme, pending the outcome of proceedings to recover the unduly paid grant from the beneficiaries, and deducting the expenditure from the next statement of expenditure only once recovery has been effected. As follows from the quoted legal provision, an error is considered corrected for the purposes of this guidance when the irregular amount has been deducted (via a withdrawal or a recovery) from a statement of expenditure submitted to the Commission or when the expenditure at stake has been registered as a pending recovery in the Certifying Authority's accounting system. 5.2. Corrective measures as subsequent events Based on the analysis of the results of the systems audits and of the audits on operations reported by the AA, the Member State (Managing or Certifying authority in accordance with the MCS) needs to take the necessary corrective, follow-up measures. If these measures are implemented before the ACR is submitted to the Commission and the AA has sufficient evidence on their effective implementation, they should be reported in the ACR, to demonstrate that the detected errors have been appropriately followed-up by the national authorities. Such corrective measures can be considered as subsequent events which have occurred after the audit period and which AA can take into account when 14/23
establishing the level of assurance and audit opinion. In coherence with the previous guidance on ACR 10, "some subsequent events might have an important impact on the functioning of management and control systems and/or on the qualifications (in cases of qualified or adverse opinion) and therefore cannot be ignored by the audit authority". These events may correspond either to positive actions (e.g. corrective measures implemented after the audit period) or have negative impact (e.g. deficiencies in the system or errors detected after the audit period). The AA may disclose an unqualified opinion if the corrective measures taken result that the risk of material deficiencies in the management and control system has been mitigated to an appropriate level, i.e. the amount at risk in the expenditure declared in year N, after the corrective measures are implemented, is not higher than 2% of that expenditure (except for a MCS initially assessed as category 4, see section 2.7). If the corrective measures concern correction of irregular expenditure, such corrections can only be considered by the AA for the purposes mentioned in the previous paragraph if the related expenditure has been correction in accordance with section 5.1. The concept of provisional withholding of expenditure by the Certifying authority is not a corrective measure in the above context. Nevertheless, the AA may consider such provisional withholding of expenditure as a subsequent event if such measure is taken before the ACR is submitted to the Commission and the CA makes an explicit commitment in writing (e.g. in a letter to the Commission) that it will only declare the expenditure at stake after obtaining reasonable assurance on the respective legality and regularity and after informing the Commission of the actions taken to obtain such assurance and an opinion from the AA on the adequacy of those actions. If the corrective measures concern an action plan, this can only be considered by the AA for the purposes above-mentioned if the actions have been effectively implemented and the AA has clear evidence thereof. In any case, the total projected error rate should remain the same, i.e. it is not affected by such corrective measures and should be reported in the ACR. 5.3. Option 1: extrapolated financial correction The Member State (the Managing Authority or the Certifying Authority in accordance with the MCS), after obtaining the total projected error rate (as disclosed in the ACR) may decide to eliminate irregular expenditure declared. This can be done by applying an extrapolated financial correction to the total unaudited expenditure of the operational programme for that year N. The above approach is up to the decision of the Member State, it is not mandatory. 10 See section 8 of the Guidance on ACRs and Opinions. 15/23
After correction of the errors in the sample, the extrapolated correction is to be applied to the whole unaudited population: extrapolated financial correction = projected error rate * expenditure not audited. 5.4. Option 2: Correction of each type of error All errors should be corrected, including known and anomalous errors. Concerning systemic errors, the AA should confirm for the purpose of the ACR that: The total amount of expenditure declared to the Commission affected by those systemic errors is determined and the responsible authorities proceed to the necessary correction 11 as soon as possible. The delimitation of the systemic error in the non-audited expenditure can be performed by the MA under the supervision of the AA. In practice, this would mean that the AA would have to review the quality of the MA's work and provide explicit, written confirmation to the Commission that the work has been carried out to the appropriate standard and that the conclusions are agreed. In order to mitigate the risk of material errors in future declarations of expenditure, the responsible national authorities should commit to implement a remedial action plan with strict deadlines addressing the systemic deficiencies. The action plan should be described clearly and concisely in the ACR. Random errors can either be the sole source of error identified in the audited sample or exist in addition to systemic errors (identified and treated as described above). As stated in section 2.3 of this guidance, the concept of random error presumes the probability that these errors are also present in the non-audited expenditure. As a result, the AA should calculate the expenditure at risk by applying the projected error rate (relating to the random errors found in the sample of operations audited) to the expenditure not audited, after deducting the expenditure affected by systemic errors. The amount at risk in the expenditure declared in year N should be disclosed in the ACR and appropriately followed-up by the Member State. 5.5. Net off with an expenditure "buffer" It may happen that in year N+1 the Member State declares to the Commission in relation to a given operation more expenditure than the one that was initially budgeted in year N. This corresponds to an expenditure "buffer". For example, a project's budgeted expenditure is 100 keur, and the public cofinancing is 40% of expenditure, up to a maximum 40 keur. The project has declared expenditure of 110 keur and received the maximum grant of 40 keur. The AA audits the 110 keur declared by the project and identifies ineligible expenditure of 9 keur. As a consequence, the national authorities may not issue a recovery order to the beneficiary because there is still enough eligible expenditure to entitle him to the maximum grant of 40 keur. 11 See Article 98(4) of regulation (EC) N 1083/2006 16/23
Another example is when a project's budgeted expenditure is 100 keur, and the public co-financing is 40% of expenditure, up to a maximum 40 keur. The project has declared expenditure of 110 keur. The AA audits the 110 keur declared by the project and identifies ineligible expenditure of 15 keur. As a consequence, the national authorities should issue a recovery order for 5kEUR to the beneficiary because the error exceeded the trigger point (100kEUR). The beneficiary is entitled to a maximum grant of 38kEUR (95*40%). 6. IMPLICATIONS FOR SAMPLING RESULTING FROM HIGH ERROR RATE A high total projected error rate may be an indication that the assumptions used when planning the sampling were not correct, e.g. the expected error rate is too low or the confidence level obtained from control testing is too high. ISA n 530 (see point 5 of its appendix 3) reflects this view in the following terms: "The greater the amount of misstatement the auditor expects to find in the population, the larger the sample size needs to be in order to make a reasonable estimate of the actual amount of misstatement in the population. Factors relevant to the auditor s consideration of the expected misstatement amount include the extent to which item values are determined subjectively, the results of risk assessment procedures, the results of tests of control, the results of audit procedures applied in prior periods, and the results of other substantive procedures." Future sampling should take account of more appropriate parameters from the experience gained Moreover, if the AA uses the MUS for the selection of operations to be audited, it would normally use a low expected error rate (e.g. 10% of the materiality level or 0,2% if the materiality is set at 2% 12 ). However, if the respective sampling results show an projected error rate above 2%, the AA may consider using a different sampling method, as reflected in section 6 of the Commission's "Guidance note on sampling methods, namely in the table reproduced below 13. Population variability Low High Low High Expected error frequency Low Low High High Suggested approach Variable sampling Monetary unit sampling Monetary unit sampling Variable sampling Clustering or stratification (plus appropriate sampling methods) 12 13 As stated in Annex 4 of Regulation (EC) N 1828/2006, the maximum materiality level is 2 %. The AA may chose to define a lower percentage. See section 6 of the mentioned guidance. 17/23
7. MOST LIKELY ERROR AND UPPER ERROR LIMIT As set out in the guidance on sampling methods, the lower error limit and the upper error limit should also be calculated and disclosed in the ACR, in addition to the sample error and the MLE. The Commission's guidance note on sampling methods for Audit Authorities states 14 the following in regard to the evaluation and projection of errors resulting from a sample selected using MUS: «An upper misstatement limit should be calculated as the sum of the projected misstatements, the basic precision (...) and an incremental allowance for widening the precision gap. The auditor can also calculate an additional sample size needed by substituting the most likely misstatement from the sample evaluation for the original expected misstatement in the sample interval formula and determine the interval and total sample size based on the new expectations. The number of additional sample items can be determined by subtracting the original sample size from the new sample size. The new sampling interval can be used for the selection. Items should be selected that are not already included in the sample. For instance, if we observe a single misstatement of 300 (25%), i.e. a projected misstatement of 1,000, with a tolerable misstatement of 5,000 and a MUS step of 4,000 at a 95% confidence level (confidence factor 3), we have a total of 13,750 of upper misstatement limit. This figure is the sum of: the projected misstatement of 1,000, the basic precision of 4,000 x 3 = 12,000 and the allowance of (4.75-3-1) x 1,000=750 (4.75 is the reliability factor (RF) for 1 misstatement at 95% confidence level, 3 is the RF for 0 misstatements at 95%). This upper limit is greater than the tolerable misstatement; hence we conclude that the population misstatement is above the materiality threshold. We also conclude that we are 95% sure that the population misstatement is at most 13,750. When applying a statistical method, the audit authority will estimate the most likely misstatement in the population and compare this to materiality in order to evaluate the results.» 14 See page 32 of the Guidance note on sampling methods for Audit Authorities for the first three bullet points; page 40 for the last bullet point. 18/23
When the MUS is used to select the sample of operations to be audited, the Most Likely Error (MLE) is considered to be the projected random error rate. This error rate forms part of the total projected error rate which is to be compared against the materiality level of 2%. When all errors found in the sample are considered as random, the total projected error rate is the random projected error rate. In any case, the analysis of the Upper Error Limit (UEL) still needs to be done, in order to respect MUS requirements. The INTOSAI guidelines 15 «represent a common base that can be referred to (...) for any audits of EU activities ( ) 16. Guideline n 23 17 refers that when «the upper error limit exceeds the tolerable error but the most likely error is lower than the tolerable error» the auditor should consider: requesting the audited entity to investigate the errors/exceptions found and the potential for further errors/exceptions. This may lead to agreed adjustments in the financial statements; carrying out further testing with a view to reducing the sampling risk and thus the allowance that has to be built into the evaluation of results; using alternative audit procedures to obtain additional assurance. The AA should use its professional judgement to select one of the options indicated above and report accordingly in the ACR. 8. PARTICULAR CASES 8.1. Errors detected by AA in expenditure that was also considered irregular by the Managing Authority, Intermediate Body or Certifying Authority 8.1.1. Irregularities already detected and acted upon by the IB/MA/CA, but not yet corrected before the sample was drawn by AA As stated above, in general all irregularities found are to be taken into account for calculation of the projected error rate and reported in the ACR. This includes the irregularities detected by the AA (during its audits on operations) which have already been detected by another national body (namely the Managing Authority (MA), the Intermediate Body (IB) or Certifying Authority (CA)), before 15 16 17 See http://eca.europa.eu/portal/pls/portal/docs/1/133817.pdf See point 6 of the Technical Introduction to the INTOSAI guidelines. Applicable to audit sampling in financial audits (including examinations of legality and regularity),according to point 6.1 of the INTOSAI guideline n 23. 19/23
the sample was drawn by the AA, but have not been corrected by the Member State before submission of the ACR 18. However, if there is documentary evidence that the relevant national authorities (MA, IB or CA) have detected the irregularity and were already taking the necessary measures before the AA's sample was drawn and that the irregular amount has been corrected before submission of the ACR, such irregularity may be excluded when projecting sample errors to the population. In any case, the treatment of the irregularity concerned should be reported and explained in the ACR in the section concerning audits on operations. As a general principle, the MA should ensure that its management verifications (administrative verifications or on-the-spot checks) are carried out in a way to prevent, detect and correct irregularities before expenditure is declared to the Commission. 8.1.2. Irregularities detected during controls by IB/MA and corrected insufficiently before the sample was drawn by the AA If during an audit an AA identifies that an irregularity was previously detected during a control by another body, but the correction rate applied was lower than the correction rate that the AA considers that the IB/MA should have been applied, then the difference in the amount resulting from correction at the AA determined rate and the amount actually corrected (at the level of declaration to the Commission before the sample was drawn by the AA) is to be taken into account for calculation of the total error rate/projection of error. 8.1.3. Irregularities relating to expenditure "de-certified" after the sample was drawn by the AA After selecting the sample of operations, the AA may identify irregular expenditure in the operations to be audited that has been "de-certified" by the MS. In terms of the practical arrangements to be adopted by the AA for the on-the-spot audits, two options are envisaged: (1) In case the irregular expenditure "de-certified" concerns all the expenditure of a given operation included in the sample selected by the AA, this body is not required to audit on-the-spot such operation. The sample should not be modified, i.e. the operation at stake should not be replaced by another operation. (2) In case the irregular expenditure "de-certified" concerns only part of the expenditure of a given operation included in the sample selected by the AA, this body should audit the operation on-the-spot in order to detect if the part not de-certified is free from errors. 18 As stated above, this correction can be done by deducting the irregular expenditure (via a withdrawal or a recovery) from a statement of expenditure submitted to the Commission or by registering the expenditure at stake as a pending recovery in the Certifying Authority's accounting system. 20/23
In both cases the irregular expenditure should be taken into account in the error rate. 8.2. Net off overstatement errors against understatement errors to arrive at an overall most likely error Understatement errors should not be considered in the calculation of the sample error rate (and, consequently, in the calculation of the projected error rate) for ACR purposes. 21/23
Final version of 07/12/2011 COCOF_11-0041-01-EN ANNEX 1 TOTAL PROJECTED ERROR RATE Analysis of the errors in the sample SYSTEMIC and KNOWN errors RANDOM errors ANOMALOUS errors (exceptional) Delimitation of expenditure affected Projection in line with the "Guidance note on sampling methods for Audit Authorities" of 15/09/2008 Corrected Withdrawn Recovered Pending recovery Uncorrected YES NO Not to be included in the total projected error Amount of Systemic/known errors Amount of random Projected Errors Amount of Anomalous Errors TOTAL PROJECTED ERROR RATE = (Amounts of systemic/known+ Random projected + anomalous errors) divided by the Total Expenditure certified in year «N»
Final version of 07/12/2011 COCOF_11-0041-01-EN ANNEX 2 - TABLE FOR DECLARED EXPENDITURE AND SAMPLE AUDITS Fund Reference (CCI no) Programme Expenditure declared in reference year Expenditure in reference year audited for the random sample Amount and percentage (error rate) of irregular expenditure in random sample 19 Total projected error rate 20 Other expenditur e audited 21 Amount of irregular expenditur e in other expenditur e sample Total expenditure declared cumulatively Total expenditure audited cumulatively 22 as a percentage of total expenditure declared cumulatively Amount 23 % 24 Amount % % 19 20 21 22 23 24 Where the random sample covers more than one Fund or programme, the information on the amount and percentage (error rate) of irregular expenditure is provided for the whole sample and cannot be provided on programme/fund level. In case of non statistical sampling for small populations, the error rate of the sample. The concept of total projected error rate is explained in section 2.6 of this guidance. Expenditure from complementary sample and expenditure for random sample not in the reference year (amount) Includes both expenditure audited for the random sample and the other expenditure audited. Amount of expenditure audited. Percentage of expenditure audited in relation to expenditure declared to the Commission in the reference year.